Abstract
This work aimed to develop an efficient R(1ρ) dispersion imaging method for clinical studies of human knee cartilage at 3 T. Eight constant magnetizations (M(prep) ) were prepared by tailoring both the duration and amplitude (ω(1) ) of a fully refocused spin-lock preparation pulse. The limited M(prep) dynamic range was expanded by the measure, equivalent to that with ω(1) = ∞, from the magic angle location in the deep femoral cartilage. The developed protocol with M(prep) = 60% was demonstrated on one subject's bilateral and two subjects' unilateral asymptomatic knees. The repeatability of the proposed protocol was estimated by two repeated scans with a three-month gap for the last two subjects. The synthetic R(1ρ) and R(2) derived from R(1ρ) dispersions were compared with the published references using state-of-the-art R(1ρ) and R(2) mapping (MAPSS). The proposed protocol demonstrated good (<5%) repeatability quantified by the intra- and intersubject coefficients of variation in the femoral and tibial cartilage. The synthetic R(1ρ) (1/s) and the references were comparable in the femoral (23.0 ± 5.3 versus 24.1 ± 3.8, P = 0.67) and the tibial (29.1 ± 8.8 versus 27.1 ± 5.1, P = 0.62), but not the patellar (16.5 ± 4.9 versus 22.7 ± 1.6, P < 0.01) cartilage. The same trends were also observed for the current and the previous R(2) . In conclusion, the developed R(1ρ) dispersion imaging scheme has been revealed to be not only efficient but also robust for clinical studies of human knee cartilage at 3 T.